Uncertainty of experiment relative or absolute

[hulk78]

I calulated the time(s), then i found the uncertainty as a percentage of my results. Later on i calculated 1/time and used the uncertainty % which i originally calulated.

Could somebody tell me if it is relative or absoulte uncertainty and why?

 

[mathman]

It appears (as best as I can understand what you did) to be relative uncertainty.

 

[hulk78]

It appears (as best as I can understand what you did) to be relative uncertainty.

But could you please tell me how to know if the uncertainty is relative or absolute?

 

[mikeph]

It's relative because you calculated the uncertainty as a percentage of your results. It's absolute if you quote the error in the same units as the measurement.

I expect this would mean something like "10 seconds with 10% error" i.e. 10 +/- 1 s (absolute)

The result of inversion is "0.1 Hz with 10% error", i.e. 0.1 +/- 0.01 Hz (absolute)

 

[e.chaniotakis]

A quick and dirty way to understand what you ask is to see if the uncertainty has units. If yes it is absolute (e.g in Mikey's example
10+/- 1 sec , the error=1 is measured in seconds=>absolute).
Else, if you measure the error/mean this is the relative uncertainty -> it is dimensionless

 

[sophiecentaur]

I can understand that you don't want to get it 'wrong' but, once you see the logic of the distinction, I think you will be able to use the terms appropriately.

If your uncertainty is expressed as a "percentage" then it is Relative, by definition and, if it is given in actual units, it would be absolute.
A digital chronometer can be a couple of seconds wrong over a month. Those two figures, taken together, tell you both the relative and absolute uncertainty and which you use will depend on that actual application. You can be pretty sure that the time will not drift by more than a small fraction of a second over one hour and that might be very relevant for some measurements - the fact that it could be two seconds out would not matter if you are measuring someone's lap time.